EVALUATING QOS IN NETWORKS WITH ACTIVE QUEUE MANAGEMENT ALGORITHMS

Abstract

Packets queuing and scheduling in network routers is a key point of overall network performance. Many applications, especially applications require Quality of Services (QoS), need techniques to pass their packets throughout routers and control and/or avoid congestions in highly congested routes. Therefore, many Active Queue Management (AQM) algorithms have been developed to avoid or control congestion in routers and provide fairness among traffic flows. This paper provides an extensive evaluation performance analysis of three well-known AQM algorithms including RED, REM and traditional Drop-Tail with QoS application requirements. The evaluation performance is conducted by employing network simulator version 2 (NS2). The network performance is measured with Voice over Internet Protocol (VoIP) traffic and three performance metrics including, throughput, latency, and PSD (Probability of Sequential Drop). The analysis shows no AQM algorithm achieve all the VoIP QoS requirements, A new AQM is needed to fulfil QoS requirements and manage queue to handle unresponsive flows.